If you’re a leader and you never feel like an impostor, I have bad news: there’s a decent chance you actually are one.
Every CEO works with incomplete information. You routinely make decisions about things that people around you understand better than you do. You are often wrong. Your mistakes affect other people, and the more power you get, the less reliable the feedback about those mistakes becomes.
Complete confidence is a strange thing to aspire to. Sometimes it is just bad calibration.
The useful part of impostor syndrome has little to do with suffering or insecurity. It is the ability to act while keeping some distance between what you believe and what reality has actually proved.
This is one of the things I’m most afraid of as a founder: losing touch with reality. Losing the correspondence between the map in my head and the territory outside it.
I’ve watched founders lose it surprisingly early. Someone raises a funding round. They have not built a successful company yet; they have persuaded investors to give them money to try. Somewhere between the term sheet and the congratulatory LinkedIn posts, that modest piece of evidence becomes: I am a genius.
At 43, I’ve started to understand that some qualities that made my life much harder when I was younger may be exactly what protected me from this. I question myself a lot. When something goes wrong, my first instinct is to ask what I did. I have never been very good at turning success into a permanent theory of my own exceptionalism.
I used to think all of this needed fixing. Now I think part of it was quietly keeping me sane.
Leadership has a strange design flaw: the job requires power, and power makes it harder to stay connected to reality.
Leadership was a verb before it became a noun
We talk about leadership as if human groups naturally arrange themselves around one special person at the top. Anthropologically, that picture is too neat.
In many egalitarian societies, authority moved with the task. The person who knew where the animals were might lead the hunt; someone else knew healing or was good at settling disputes. David Graeber and David Wengrow describe societies that seem to have switched hierarchy on and off depending on the season: strong authority for a period of intense coordination, then back to a flatter arrangement when the task disappeared.
You could lead the hunt without becoming The Leader.
Christopher Boehm’s work points in the same direction. Egalitarian groups often actively restrained people who tried to convert influence into permanent dominance. Leadership existed, but authority stayed relatively close to the function that justified it.
That is still the cleanest definition I know. Leadership is a coordination job a group delegates because coordination is expensive. Somebody has to hold the shared goal, integrate information, resolve contradictions, make decisions when consensus becomes too costly, and take responsibility for whether the system works.
A group could theoretically do all of this itself, just as I could theoretically make my own shoes. I delegate shoemaking because somebody else can do it better and I have other things to do. Leadership is much closer to that than to kingship.
The group does not exist as an audience for the leader. The leader exists to make the group more capable of reaching the goal.
“I’m the founder here, goddammit”
At Welltory, we have an internal joke for the moment someone forgets this.
One of our employees once told us about a company where a serious argument had dragged on for a long time. Eventually the founder lost patience, slammed his fist down and said:
“I’m the founder here, goddammit.”
Apparently, that was the exact moment several people in the room lost respect for him. I understand why. If, in an argument about what is true or what the company should do, your final piece of evidence is your place in the hierarchy, you have admitted that you ran out of evidence.
So now I sometimes say it as a joke when a debate at Welltory has gone on forever. Everyone laughs, and then I still have to explain why I think I’m right.
Being the founder may give me the formal right to make the final call. It does not make the call correct.
Founders are the people who know where the mammoths are
Founders are a very good example of why functional leadership exists.
A founder starts with something like: I think the mammoths are over there.
I see a market nobody else sees. I think this product should exist. I think these people can build it. If others believe there is a reasonable chance you are right, giving you more authority makes sense. You hold the original model, know context nobody else knows yet, and probably took more risk.
Authorship can be a legitimate source of epistemic authority. The trouble starts when that authority expands beyond its evidence: I know where the mammoths are, so I must be unusually good at finding mammoths; perhaps my judgment is unusually good in general; therefore I should make the decisions.
A few steps later, a specific achievement has turned into a personality theory.
Founders seem susceptible to this before success even has time to do much damage. In a classic study of almost 3,000 new business owners, 81% rated their own chance of success at seven out of ten or higher, and one-third said ten out of ten. The same people gave lower odds to other businesses like theirs.
In other words: businesses like mine are risky, but somehow the risk applies less to me.
Founder skill is real. Successful entrepreneurs do have a higher chance of succeeding again — around 30%, versus roughly 22% for first-timers. That should update your estimate of your own ability. It should not turn 30% into divine appointment.
The dangerous conversion is from I was right about this to I am the kind of person who is right.
How authority became something you could keep
Graeber and Wengrow destroy the cartoon history where humans were egalitarian until agriculture arrived and immediately invented kings. Humans experimented with hierarchy long before farming. What some forms of agriculture changed was the economics of making power permanent.
Mayshar, Moav and Pascali found that hierarchy was associated not simply with productive agriculture, but with crops that were easy to appropriate. Grain ripens predictably, sits above ground, can be counted, stored and seized. Root crops can stay underground.
Once a group produces a surplus someone else can reliably take, coercion can finance itself — and authority becomes much harder to revoke. If a hunting party no longer trusts the person leading the hunt, they can stop following him. If the person claiming authority controls armed people and can take your grain, replacing him is harder.
So maybe the important political invention was authority that could survive after the function that justified it had disappeared.
This still matters in companies. Leadership is authority attached to a useful function. Power can continue long after that function is being performed badly.
Power amplifies what it finds
The standard phrase is “power corrupts,” but a lot of psychology on power is better described as disinhibition. External constraints matter less. People become more approach-oriented, and whatever motives, preferences and traits were already inside them get more room to drive behavior.
Power makes you more able to be yourself.
That can be excellent news. If what you deeply want is to build something useful, help people or accomplish a shared mission, power lets you scale that motive. If what you deeply want is status, admiration, domination or proof that you are superior, power scales that too.
Then correction gets worse. People edit themselves around you. Your ideas get more resources because they are yours. Bad news travels slower. You depend less on other people and therefore have fewer automatic reasons to model them carefully.
There are even neuroscience findings in this direction: experiments that induce a sense of high power have found weaker automatic motor resonance with another person’s actions, alongside behavioral findings of poorer perspective-taking in some contexts. Power does not appear to produce literal brain damage; it can make the system less automatically tuned to other people while making action and internal goals more dominant.
For leadership, that is a nasty combination: more freedom to act out your motives while the world becomes worse at correcting them.
Why some people can never have enough
Why does someone with a billion dollars need ten billion? At some point, money stops meaning comfort, safety or freedom in any normal material sense.
Unless money has become a score.
Psychologists call one version of this contingent self-worth: your sense of value depends on achieving something external. A lot of male socialization seems to attach worth to winning — status, competence, wealth, toughness, dominance. The precarious-manhood research adds another layer: masculinity is often experienced as something publicly earned and therefore publicly losable. Threaten it, and some men compensate through status-seeking, risk-taking or dominance.
If your sense of worth is tied to winning, success never finishes the job. It gives temporary validation and then moves the reference point. There is always someone richer, more famous, more powerful, or running a more valuable company.
The loop feeds itself: self-worth depends on status, status produces validation, validation fades, and the next marker becomes necessary. Power makes this especially dangerous because it amplifies the status motive while removing some of the people who might interrupt it.
There is a basic difference between someone who wants power to do something and someone who needs to keep doing things in order to feel powerful. The first person can eventually have enough. The second has no natural stopping point.
Women get a different bad deal
Girls are usually handed a different set of contingencies. The message is less “you matter if you win” and more “you matter if you are valued.” Be liked. Be attractive. Be useful. Maintain relationships. Notice what everyone else needs. Don’t be too aggressive. Don’t take too much space.
This produces a different set of leadership liabilities: people-pleasing, conflict avoidance, overpreparing, treating other people’s emotions as your responsibility, waiting for proof long after someone else has already claimed the role on potential.
We can see pieces of this in the data. In a study of almost 30,000 employees on a management track, women received higher performance ratings and lower potential ratings. The potential gap explained a substantial share of the promotion gap, even though women with the same potential ratings subsequently performed better.
Performance is evidence. Potential is credit.
Women get asked for more receipts and extended less credit. At equal objective performance, women also describe their own performance less favorably than men do. Meta-analytic work on perceived leadership effectiveness finds a similar split: other people do not rate men as better leaders, while men tend to rate themselves more highly.
All of this is expensive on the way up. Then the sign can flip.
The sign flips
Before power, the environment tends to discount women. After power, the environment tends to inflate leaders.
A habit that was costly on the way up can become protective once the distortion reverses. Nobody has run a 30-year randomized trial on this, obviously. This is how the pieces fit together from where I sit.
A woman spends decades learning to read other people, monitor her own behavior, gather evidence before claiming authority, and wonder whether she might be wrong. Then she gets authority and enters an environment where people hesitate to contradict her, her ideas get resources because they are hers, and success becomes easier to attribute to her skill than to luck.
At that point, the useful residue of all that self-doubt is specific: your own wrongness still feels like a live possibility.
That is surprisingly valuable once everyone around you has reasons to tell you how right you are.
Midlife may make this more interesting
This became especially interesting to me when I realized how much I had changed by 43.
Women do not stay psychologically identical from 25 to 45. Longitudinal work finds that through midlife women on average become less dependent and self-critical and more confident and decisive. Self-esteem rises through much of adulthood, while body surveillance and body shame tend to decline.
There is no neat hormonal switch called “menopause makes you a better CEO.” Age, perimenopause, accumulated competence, financial independence, older children, changing reproductive incentives and plain exhaustion with caring what everybody thinks may all arrive in the same general period.
The combination is what interests me. Women may spend the first half of adulthood developing very sensitive social sensors: reading people, monitoring reactions, noticing what they missed. Then part of the need to obey those signals starts weakening.
You keep the sensors and lose some of the obedience.
That may be an unusually good leadership configuration: high sensitivity to other people with lower dependence on their approval.
And this is where the trajectories can diverge. For a man whose self-worth has become attached to winning, success can strengthen the old contingency. Winning proves worth, so more winning becomes necessary. More status produces more power, power disinhibits the motive, and weaker feedback makes the self-story easier to maintain.
For a woman, midlife can sometimes produce the opposite loop. Real competence creates evidence-based confidence. Lower approval dependence makes difficult decisions easier. Social sensing and self-questioning remain. Confidence grows from results rather than rank.
One loop can carry you farther from the territory while making you feel increasingly certain. The other can make you more confident while keeping reality in the room.
That pattern suddenly explains a lot of what I’ve seen, including in myself. I am intellectually much more confident now than I was at 25. Emotionally, there is still a piece of me that refuses to fully believe my own press release. I used to want to get rid of that mismatch. I don’t anymore.
Confidence is a terrible KPI
Calibration is much more useful.
In one study of more than 2,000 managers, people who overrated themselves relative to how others rated them were judged less effective than both accurate self-raters and people who underrated themselves. I would not build a religion around one study, but the asymmetry makes sense.
If I underestimate myself, I may lose an opportunity. If I overestimate myself while running a company, hundreds of people can end up executing my mistake.
So by all means, get rid of the anxiety and the part of impostor syndrome that makes you miserable. Keep the capacity to wonder whether the evidence supports you. You need enough confidence to act; becoming immune to correction is not the upgrade.
Responsibility starts with you
For me, this comes back to the simplest definition of leadership: responsibility for the outcome of the system.
When something seriously goes wrong at Welltory, my first question is usually: what did I do that caused or allowed this?
That does not mean every failure is personally my fault. It means every important failure contains information about the system I am responsible for. Which incentive did I create? What did I tolerate? Which signal did I miss? Why did the information reach me this late?
This is why concentrated authority without concentrated responsibility is intellectually dishonest. You cannot say, “I should control this company because I understand it better than anyone,” and then explain every repeated failure as “my VP fucked up.” Maybe the VP did fuck up. But if your system keeps generating the same kind of failure, that eventually becomes information about you.
The more power you keep, the more aggressively reality should be allowed to update your model.
Don’t ask for feedback. Build sensors.
Leadership advice loves the idea that if you create enough psychological safety, people will tell you the truth. I don’t trust this as a primary feedback system.
Once you have real power over someone, you have changed the measurement environment. If their salary, promotion, project or job depends on you, they cannot give you exactly the same signal they would give a peer. The distortion runs both ways: people stop telling you when you are wrong, but they also stop telling you when you are unusually good at something because praising the boss feels sycophantic or unnecessary.
So I don’t want my sanity to depend on everyone around me overcoming their incentives. I want instrumentation.
Where reality can be measured directly, measure it. Build alerts, analytics and monitoring that reach you before somebody has to make a political decision about whether to tell you. And keep a record of your own judgment.
I now use AI to automatically capture decisions and predictions I make in meetings, then compare them later with what actually happened. Six months later I don’t get to say, “I always knew this was risky.”, if the record says I gave it an 80% chance.
Over a few years, that becomes a personal calibration dataset. When I say 80%, am I right about 80% of the time? Which decisions am I systematically overconfident about? Where am I too pessimistic?
Then there is hiring. I want people around me who can disagree properly. A professional contrarian who turns every disagreement into a six-hour war of attrition is not a safeguard; that person is just exhausting. The valuable person can challenge my reasoning clearly before the decision and still commit to executing once the decision is made, even when I chose differently.
And I need people outside the company’s incentive structure too: a therapist, a good coach, another founder who does not work for me. Someone with whom my title has no exchange rate. It’s really hard. I only managed to have a therapist and a coach I talk to once a year. I've never been able to find a mentor or founder whose experience is relevant.
So I increasingly think the answer is to build systems that rely less on spontaneous honesty: log your bets before hindsight rewrites them, instrument the company, and hire people smart enough to tell you you’re wrong without trying to take your chair.
Find the mammoths. Then keep checking.
Founder control is not the problem. The dangerous founder is the one who found the mammoths once and concluded that this makes them the permanent universal authority on mammoths.
Knowing where to go can legitimately make you a leader. The job after that is to keep the group connected to reality, including when reality proves you wrong.
So if you still occasionally wonder whether you are an impostor, I would not rush to cure every last trace of that feeling.
Just keep checking where the mammoths actually are.